I was given a complete, very detailed model of a Ford 427 SOHC Engine (aka "427 Cammer")pretty much complete down to the last bolt and washer, full scale, and dimensionally accurate for everything I have checked. Amazing! The big downer was it was done in SketchUp which is a surface modeler instead of parametric solid modeling software. Some of you may have seen my other post looking for a SKP aficionado. A friend who is a model engine builder asked me if I could/would make a set of engine castings. It did sound kind of fun. To cut to the chase, I was able to use a free trial of Sketchup to repair and export clean STLs of all the parts that would be engine castings. The part with the finest detail was the valve cover. I wanted to see the art of the possible with the level of detail possible for my lost foam processes. I made a quick program for my CNC Router to make a sample pattern. Each side was about 8 minutes with a .125"D ball end mill and .1 step over. I chased the ribbed and logo/lettered area with a .031"D ball end mill which barely fit the letter spacing and as you can see, did not clear all of the logo features. Here was the result: My take: Not good enough. These would need to be lost resin/block investment cast probably with Zamak to get the detail. If you have to resort to < .031 cutter for detail you're probably not going to get there machining foam. Some of the other parts like the intake manifold, timing cover, bell housing, block, heads, etc may be possible in LF but fit and finish would be better as printed pattern and block investment cast, but they could probably be done in aluminum. Over time I may take it on just as a test bed to develop more capability, but buying a 3D printer, building a vacuum investment casting rig, and going through the learning curve for both is a big ask for a favor. I have so many full size automotive castings on the go it's hard to justify but it's a nice diversion and daydream. -Maybe over time. Anyone here done any 3D printing with any of the evaporative casting resins? Best, Kelly
Yes. The sizes of the U-joints in the photos are the machine screw size for the pin and bolts. The rectangular plate is 1.6" long. The amount of detail that one can get is unbelievable. Your valve cover would be about 8" long. It could be done as an investment casting, but would take a fairly big flask, quit a bit of investment, and a fairly large burn out oven. A "wax" that large could only be printed on the high end (size wise) of home 3D printers.
I've done all my investment casting in bronze. I've seen suggestions that it could be done in aluminum, but haven't seen evidence of anyone doing it successfully on any videos or on web sites. I think investment castings could be done in zinc alloy. I've sand cast ZA-27 in Petrobond and it's nice to cast. However, for precision castings in zinc I think injection molds are the norm. I wonder if you could make a hybrid of foam with a wax insert? I don't know that the resins would work as I don't think they melt, they vaporize. Another possibility would be an insert of an already burned out piece of investment which had voids where the letters would be, glued to the foam.
Something I've wanted to try is to use a 3D printed "wax" pattern and use sodium silicate as a binder for a fine sand. Then burn out the pattern in the normal way. The sodium silicate should withstand the temperature needed to burn out the pattern. It's my understanding that sodium silicate can be set with heat, so CO2 may not be needed. The mold could be poured at an elevated temperature to help the metal to flow and encourage it to fill the mold. I'd like to try both gravity feed and vacuum assist. That might be a way to pour your valve cover without having to invest in a large amount of investment power. I guess this would actually be a variation on shell molding. Has anyone tried investment casting using sodium silicate bonded molds?
Thanks for those replies Dick. I remember that post and those u-joints are impressive indeed. Just to clarify, I certainly know it (evaporative 3D resin prints for block investment casting) could be done. I was looking for someone to print me as sample in a casting resin. I make small highly detailed wax appliqués all the time and glue them to my foam patterns. The refractory coating reproduces the detail very well similar to good investment. It works great. Here's an example. I had a little learning experience with the silicone reacting with the print resin but eventually succeeded. https://forums.thehomefoundry.org/index.php?threads/two-barrel-carburetor.947/page-4#post-23586 But you might be surprised at the level of detail possible with foam patterns: Testing the Limits of Feature Size | The Home Foundry The problem with these parts is at their size they're sort of tweeners. The block is approximately 6"x 4"x 4", the pattern and mold materials would get to be a bit pricey. A pair of heads, and a pair of valve covers could be done in the same size flask. The pattern price would recur for every part. In addition to quality investment, the evaporative casting resins would be fairly expensive for that size part and yes, all though manageable, the print times are longer. If someone isn't experienced with the particular resin, there could be some trial and error and it gets to be a bigger ask to print a couple. For the valve cover, it could be done with a conventional printed resin and then use the traditional lost wax method, making silicone molds, waxes, burn out, etc..........but, that is a lot of work, especially compared to lost foam. I think getting the valve cover detail in Aluminum would be very difficult but in a zinc/Zamak alloy with vacuum or pressure vacuum block investment, no problem, as long as you had the right investment and could get the mold clean. The block and heads have much better chance of success in aluminum. The problem with machining common XPS foam at this level of detail is the ornamental dimensional features start to approach the foam cell size so you get tear out and poor as-machined finishes whereas on my full-size automotive parts it's virtually unnoticeable. I may be able to get the desired detail with a machinable wax. Ideally the block and the heads would have internal water jackets. Cores in the full-sized parts are considered fragile. Imagine what that means for these parts. I actually think there is the possibility the heads and block could be done in lost foam. The surface detail is much less on those parts. Here is an example why. This was 8+ years ago and my processes are so much better now. Unsupported Closed Loose Sand Coring In Lost Foam | The Home Foundry The reason I cling to lost foam are many, but there is no burn out required, and it only takes minutes to machine a pattern of this size not hours or days, and pattern/mold costs are essentially zero. Once I have a coated pattern I can usually have a casting an hour later. MUCH less laborious than traditional lost wax. My foundry furnaces are essentially PiD controlled kilns so burn out block investment is no problem other than cost, time, and odor. I'll be waiting for ESC to weigh in here. He did his V8 with traditional sand casting methods.......bless his heart! Best, Kelly
Here I am. Remember this was a hobby. I have about 4000 hrs in this build. I started it in 1997 and I don't remember quite what prompted me to use 1/3 scale but I think it was the fact that the aluminum cast Challenger kit was approximately that scale was the reason. It was a flathead on the Lycoming , Cord pattern with the exhaust in the valley. It had a flat crank and a 1" bore and stroke and although a decent runner still struggled to overcome the friction to energy balance. So I went for cubic inches to overcome that. The 1/4 scale engines suffer more and are not cast , but EDM or CNC machined although some of the recent miniatures seem to run well. I spent almost two years on the heads for the smallblock. I think I have pictures of the interlocking waterjacket, intakes and exhaust cores mounted on a follower. I'll see if I can dig some up. All the cores are simple draft oil bonded core sand, individually vented and glued with flour paste. As I remember a pair take about 16 hours including the bake time. All this to say I prefer 1/3 scale for casting and then this might be doable in foam. Using your wax for detail and leaving interior material that could require machining for clearance.
Kelly, Thanks for the response. You referenced several topics that you posted before I was a member. All interesting information. I've never tried lost foam and probably never will, about 95% of what I know about LF is from your posts. The individual cost for the "waxes" would be nominal. The resin I use costs $70 per KG. Just guessing, the resin to print a pattern for the valve covers would only be in $3-$4 range. The valve cover is smaller than I thought and within my printer capacity. It you share the stl I can see if I can print it. Unfortunately, I can't offer to cast the part. A problem for me with casting large parts is that I live in Alaska and most suppliers of the investment powder either won't ship to Alaska or will only ship second day air so the price quickly gets out of hand and I stick with smaller parts. I also don't have a flask big enough for the valve cover. It's disappointing because I'd like to play with investment casting aluminum or a zinc alloy. All my investment casting has been in bronze for parts for a 1/8 scale steam locomotive (and a few small parts for a full-sized steam locomotive) where there is no call for aluminum parts.
That's very generous of you Dick. There is one problem with printing the Valve Cover from the files I currently have. When it was modeled, the thickness was only .062" wall full scale. Consequently it is only .015" wall when reduced to 1/4 scale. Not sure I have a way to edit the STL file to reduce the size of the interior to increase the wall to something more sensible for casting. When I machine the interior, I'm able to command a roughing clearance when machining the interior and that results in about a .090 wall (or whatever I want). Does your printing/slicing or other software have such capability? Best, Kelly
I decided to machine patterns for some of the other "easier" engine parts. Below are the front timing cover, bell housing, and oil pan shown along with the previously machined valve cover. They are all about .090" wall. That's quite sporty and will be quite the challenge to cast. Best, Kelly
My 3D cad software (Freecad) can import an stl file and convert it to a format that can be edited. Unfortunately, that's something I've never tried. It will also offset a surface. Something else I have never tried. These are skills that would be very useful. If I have some spare time maybe I'll try experimenting with those functions.
No draft inside or outside. Ok for machining and printing evaporative patterns but might get a bit difficult to extract waxes from silicon molds depending upon the part. I'd say sand or any reusable pattern hard mold method would be a no-go without adding draft. Best, Kelly
It isn't something I have experience with, but another possibility is machinable wax. A Google search for "machinable wax for investment casting" brought up some information, including that it is available from online metals dot com and isn't crazy expensive. What about a machined wax insert for the fins and logo with the rest being foam? That way you could go with the equipment you have.
I've use machinable wax before but it is more like the (higher) density of investment wax and would likely create some problems in the lost foam process and may need to be burned out. The wax I use for lost foam casting has the same melt point as the foam. I think I'll get there with the foam. I made some further improvements on the Valve Cover details. There's actually a left and right valve cover. Instead of allowing the 3D MOP on the STL surface to cut the artwork, I isolated and filled the graphics area by generating a patching solid used along with the STL, and edge detected the graphics from the original STL so I had them as polylines. I enlarged those graphics polylines by 10% to give me a little more operating space and then proceeded to make subtle changes to the fonts by dragging the nodes/handles on those polylines to open the pinch points and increase font width in areas where it was too thin to survive cutting. In most cases it was only a few thousandths to allow the cutting diameter to clear the feature. With the graphics being polylines, I could now cut them with a Pocket MOP which is much faster and even with a .5mm EM the run time was only 60 seconds for the logo. I could make additional drag/shape changes to the polylines without changing the pocket program, so being < 1 min/cycle, I started experimenting. Here's where I got to. I sectioned one to take a look at the wall thickness and it looked reasonable. Still not all the way there but close enough to see what is possible and needed to cast them. In the close up you can see I'm starting to bump into the cell size limitations of the foam and some micro tear-out. I might use higher density foam or try coating/filling with paste wax by painting thee details with a small art brush. These small patterns are kind of fun. With the short cycle times I can blow them out at about 15 minutes/copy. Since I already had all the geometry from the parts I used it to create ring gates (sitting next to the timing cover and bell housing in foreground and sitting on their respective gating in the background) that distribute the hot metal from a thick runner all around the perimeter of the casting. The rest of the parts don't have the graphics details like the valve covers. I dipped coated four valve cover patterns. Supposed to come out of the deep freeze and have some better casting whether in the near future. Did I mention they are thin walled!! Best, Kelly
Did some casting development work on the valve covers today. First attempt I just poured like I would one of my normal castings, 1350F with gravity feed. It was a fail and not even close. Couldn't even fill the gating system and barely penetrated the pattern. The first two patterns were double dipped and had pretty thick refractory coating. So then I went to the other end of the spectrum, 1600F pour temp and applied 1/2ATM of vacuum to the flask, and viola. So then I cast two more that way. The first two patterns were double dipped and had pretty thick refractory coating. The last two were single dipped with a second coat brushed applied only to the ribs and graphic's. Here are some close ups of the last two: Here's a picture of my vacuum boost for the flask and links to my builds for the vacuum rig and flask: Vacuum/Forced Air Cart | The Home Foundry My New Lost Foam Casting Rig | The Home Foundry I can make some little scrapers to clean up the rib valleys and graphics but I still have a few more tricks up my sleeve to improve the pattern quality. This gives me great confidence for casting the other parts which are of similar size and wall thickness but not as fine detail. -Definitely pushing the limits of lost foam casting here........but that's what I do! Here's a short video of the results: Best, Kelly
Great job Kelly. I thought about offering to try one as a pattern, but even if I fixed the draft, pulling the fins and Logo would have been a bear. Somewhere there is a running 1/4 scale? hemi that was machined from the solid with a water pump I cast from the plastic model kit. There were no cores so he drilled the coolant passages and fed one way.
Thanks ESC. I think it would be an exercise in futility with the Ford logo. Maybe with some very fine bound media, but even then, releasing without breakage.....as you say, a real bear. Believe it or not, I do have a router bit that is 1/32"D ball nose with 3-degree (IIRC) taper so I could probably machine a hard pattern with draft from the CAD/STL files I have......but it would take some doing. I do have the water and fuel pump housings too. They're at the bottom of the priority list at the moment. Block and heads come before them. They will likely need to be cast solid. The discharge hose barb is only .5" OD and with any appreciable wall thickness the ID would be almost zero. The impeller cavity exterior features could be net shape. Best, Kelly
I will see if I have one of the Mopar water pumps in my stash. The builder routed the coolant through the most direct leg to simplify the machining. Then through the first head/block and then across via the manifold to the second head/block and finally out through the intake return. Since it was a tabletop runner there was no issue with overheating. I must have missed the introduction for this project. Are you going to build a runner?
I showed a friend who is (or at least has) a model engine builder the CAD model and after some discussion he asked if I'd make him a set of castings. I'll keep a set for myself but I'm not planning on building a running engine myself in the foreseeable future......but never say never. Everybody wants a valve cover LoL! Best, Kelly
